The race is on to make buildings not only more efficient but truly intelligent. It’s a tricky path full of hurdles and breakthroughs, as companies try to blend cutting-edge tech into the bricks and mortar of our daily lives.
Enter Johnson Controls, a company focused on enhancing building functionality. They’re mixing old-school building know-how with the newest digital tricks to push the envelope on what a building can do.
In an interview with Frontier Enterprise, Vijay Sankaran, Chief Technology Officer at Johnson Controls, explores the integration of AI, the Internet of Things (IoT), and new tech trends. He discusses how these advancements are influencing smart building management and the strategies Johnson Controls is deploying to navigate industry changes.
In practical terms, how has AI implementation in building management systems directly affected operational costs for your enterprise clients?
The impact of AI on operational efficiency in buildings is significant across our client base, particularly in energy optimisation, achieving net-zero compliance, and enhancing space utilisation for improved comfort and efficiency. AI underpins these improvements, enabling a comprehensive approach to building management.
In areas such as energy efficiency, collecting data on how various equipment utilises energy and affects heating and cooling allows AI to optimise these processes. This optimisation is based on the specific utilisation of spaces within the building or outside temperatures, and how they may need to vary the settings.
Coupling AI with the actual command and control of the site-based systems and varying the points can drive efficiency from a building perspective. Similarly, for net zero, AI allows you to distil scope 1 and 2 emissions for a site and forecast net-zero compliance. You can also use AI to facilitate scenario-based planning for energy conservation.
Across Asia-Pacific, many real estate managers and owners are turning to AI to drive energy efficiency improvements. This is achieved by assessing, benchmarking, and baselining a building’s energy efficiency, then employing automation and autonomy to adjust settings in various spaces, thereby leveraging AI for significant energy savings.
We’re just at the tip of the iceberg in terms of AI. I also see generative AI playing a large role going forward, especially in generating specific recommendations for building operators to enhance efficiency. We’re right at the tip of the spear in terms of what AI is driving from a smart building standpoint.
What are the biggest technical challenges that you faced in integrating the AI into existing building management systems? How did you overcome them?
The biggest challenge in utilising AI within building management is twofold: data availability and system integration. Initially, the lack of sensor or BMS data in many spaces hampers the effective application of AI models, such as those forecasting energy use. Without adequate metering to connect equipment with its energy consumption, accurately predicting the energy impact of individual pieces of equipment is difficult.
Furthermore, the effectiveness of AI also depends on its integration with the building management system for executing improvements. That presents a challenge with older or obsolete systems. This integration is essential for leveraging AI to deliver tangible benefits through closed-loop systems.
Ultimately, the key to overcoming these challenges is in ensuring access to comprehensive data, which is crucial for deploying AI models that accurately enhance building operations and create value.
From a technology perspective. How does retrofitting older buildings differ from integrating tech in new constructions?
Retrofits represent a bigger opportunity for energy efficiency and decarbonisation, more so than newer buildings, which already incorporate the latest, modern equipment and control systems, including high-efficiency lighting. This makes achieving substantial energy efficiency improvements in new facilities more challenging. The real value in newer facilities often lies in how effectively they utilise space based on occupancy patterns. Space utilisation is really important.
Deploying new spaces requires considering them as dynamic, with people constantly moving in and out. For instance, this conference room I’m in will be vacant post-interview. We probably don’t have the means to control this room’s temperature and change it by a few degrees when it’s not occupied. This situation often stems from not having sensors to monitor room entry and exit, alongside the absence of specific room-level comfort controls within the building’s air handling system to adjust temperatures as needed.
To make a new building smart, it’s essential to integrate the right technologies and design the building architecture with HVAC and control systems that allow for individual area control based on the patterns and requests of the people using it. Conversely, retrofits typically allow for cost savings due to their older, inefficient equipment but lack visibility into actual operations and consumption.
Retrofits face the challenge of gathering data to establish a baseline for the energy consumption of different equipment types, often requiring manual effort. For example, obtaining real-time data from a 15-year-old chiller can be particularly challenging, especially in buildings lacking modern IoT capabilities to signal underperformance compared to more efficient spaces.
Our sustainable infrastructure practice is dedicated to older facilities, doing manually intensive benchmarking to identify energy conservation measures and advise on enhancements for greater efficiencies. This process is interesting when considering our entire customer base, as the pain points vary significantly. Retrofit-focused clients often seek basic data on consumption. In response, we frequently implement metering solutions to achieve more detailed sub-level metering. We implement customisations to gain insights into how buildings are functioning and apply turnkey solutions for space utilisation. These solutions may not be fully integrated but are designed to provide a clear visual on how sites are being utilised.
Contrastingly, smart buildings are designed from the outset with systems for dynamically managing efficiency more effectively. This presents a societal challenge, as the demand for upgrades is predominantly within the retrofit space, yet the bulk of digitisation investment targets new, high-efficiency constructions. From an open-loop perspective, our aim is to support retrofits in enhancing energy efficiency and advancing decarbonisation efforts, ensuring they can achieve these goals without solely relying on extensive retrofit or modernisation projects.
What non standard cybersecurity strategies has Johnson Controls adopted to protect its enterprise level smart buildings?
Cybersecurity is paramount for customer trust in digital platforms. With my background in financial services, I understand its importance. Recently, we acquired Tempered Networks, integrating zero-trust cybersecurity for comprehensive edge-to-cloud protection. This strategy addresses the vulnerabilities of older building assets, such as data collection servers or BMS, that might operate on outdated software, mitigating risks associated with cloud connectivity.
Zero trust creates a military-grade encryption tunnel between the edge and the cloud, ensuring that only specific individuals can access the resources within the building. It forms a completely sealed network tunnel over public internet or cellular connections, where only specific, named resources can access the information or know what devices are at the endpoints. This point-to-point connection is a significant differentiator in our approach to cybersecurity in OpenBlue and our digital solutions, addressing concerns and risks from recent attacks on operational technology assets.
We’ve also significantly invested in cloud data security, with rigorous protection and monitoring measures to safeguard essential building data. This commitment spans from software engineering to operational deployment and security at the edge, as part of our cybersecurity value proposition.
How is Johnson Controls using AI to prove to improve its own internal processes?
Across our entire organisation, we’ve established an AI council, which I co-chair with our CIO. Our CIO handles all internal corporate technology processes, while my focus is on customer-facing technology processes. Our concerted effort has led to the formation of a steering committee that looks at AI use cases in both corporate and product areas. In corporate sectors, inventory optimisation and forecasting emerge as prime areas for AI deployment, with generative AI significantly enhancing our service technicians’ ability to efficiently resolve customer issues by providing summarised guidance from manuals.
On the product side, our security business has greatly benefited from Vision AI, which has been central to our AI strategy for years. This technology is pivotal for applications such as worker safety and tailgate detection, enabling us to promptly respond to safety or security concerns. The implementation of AI spans across all facets of our operations, from the commercial side to the product domain and manufacturing and supply chain, underscoring AI’s integral role in enterprise transformation.
We start with identifying applicable use cases for AI across various processes, followed by an AI R&D phase. This phase is dedicated to testing and evaluating diverse AI solutions, determining their potential impact and readiness for broader deployment. This scrutiny is crucial, especially for managing risks associated with generative AI, assessing if the investment in AI technologies correlates with substantial organisational efficiency gains.
We’ve recognised the inherent costs associated with adopting AI, be it enterprise solutions like co-pilot systems offered by our tech partners or internal developments harnessing GPT-4. Despite the allure of these technologies, our priority remains clear: These solutions must solve every problem. The AI council’s mission is to conduct a holistic look at these opportunities, across all these different dimensions, and then figuring out which approach is best in solving the problem.
Aside from AI, and IoT, are there any emerging tech that you think will revolutionise smart building management in the next five to 10 years?
Digital twins, previously seen as a concept far ahead of its time within smart buildings, are now recognised as crucial for the future of smart building management. With the shift towards large corporations controlling commercial real estate, there’s an increasing demand for managing property portfolios more remotely, requiring less local investment for maintenance and security. This shift is driving the need for integrated command centres to facilitate more efficient operations across multiple properties.
The implementation of digital twins allows for a detailed view of a building’s status, including real-time updates on different floors and the ability to pinpoint and respond to alarms quickly. This capability supports the growing trend towards centralised management of properties owned by a single operator, marking a significant change from traditional management practices.
Moreover, the dynamics of real estate is changing, with a broader spectrum of investors showing interest in real estate funds. This change prompts real estate companies to visually demonstrate their property portfolios, leveraging digital twins not just for operational management but also as a tool to attract investment.
The concept of data as an ecosystem is gaining traction, with an increasing amount of building data being integrated into various applications. The trend towards creating data lakes for building portfolios facilitates the sharing of information through APIs for diverse purposes, reflecting a broader adoption of this approach.
Following the recent COP28 (the United Nations Climate Change Conference), the emphasis on showcasing progress towards net-zero targets using data platforms highlights the potential for government buildings to lead this effort. This scenario suggests a growing reliance on comprehensive data platforms to aggregate and share data on a national scale, contributing to a holistic view of a country’s net-zero commitments.
Finally, the integration of alternative energy models, such as grid interactive optimisation, into building management is emerging as a key trend. This approach addresses the dynamic balance between using on-site energy assets and utility-supplied power, offering a strategic framework for managing energy resources efficiently. This method, particularly relevant in regions with specific regulations like Singapore’s renewable energy certificates, is anticipated to become a significant area of growth, enhancing the sustainability of building operations in the next decade.
